Final Lectures Flashcards
2 x continuous design
2 IVs, but one is continuous (such as income on cts scale). For example, study about Europop vs rap stars’ happiness according to income. IVs are music genre and income, with income continuous and music genre nominal
Nominal vs. ordinal
Ordinal is like nominal, but there is a clear ordering of levels: “low, medium, high” for ex.
IRB approval
Institutional Review Board for Protection of Human Subjects approval necessary to conduct research
Problems with deception
Both ethical and future bias-related. Ethical: self-esteem (tricked or finding out negative things about self), no real informed consent. Bias: participants may be tainted for future studies - burned once.
Quasi-independent variables
Researcher does not truly have power over what group of IV the participant is assigned to. For ex., smokers, gender, sexuality etc. Hard to establish causality in a fool proof way due to potential of 3rd variables
Covariate
A variable that is possibly predictive of the outcome under study. A covariate may be of direct interest or it may be a confounding or interacting variable. Can try to statistically measure and remove these effects
t-test vs ANOVA
t-test compares means of groups when there are only two conditions; ANOVA can process for more.
Correlation vs. regression
coefficients are the same, however regression has a linear equation that predicts Y based on X, correlation can consist of nonlinear relationships as well
Between vs. Within vs. Mixed
Between - different groups have different participants
Within - different groups but same participants for all steps of experiment
Mixed - 1 IV is between, 1 IV is within
Pre- vs post-test data collection
Data collected may differ if you measure before or after the experiment due to practice effects/pre-test may impact later results If there is no pre-test, however, there is no baseline by which to measure the impact of the activity/treatment - there could be initial group differences.
Solomon Four Group Design
Can avoid some of the problems of pre- and post-test design, with 2 extra control groups to assess whether pre-test itself impacted subject behavior Pretest-Treatment-Posttest Pretest-Posttest Treatment-Posttest Posttest
Nested Design
Some factors are found to be naturally nested within the ‘levels’ of other factors. Between-subjects. Some treatments by their nature are nested. The effect of sex, for example, is necessarily nested. One is either a male or a female, but not both. Current religious preference is another example. Treatment conditions which rely on a pre-existing condition are sometimes called demographic or blocking factors.
Cross-sectional design
Can measure development by comparing cohorts of different ages. Quick and cheap but there may be generation/cohort effects
Longitudinal design
Measures development by examining participants over time - can avoid generation effect to an extent. Problems with subject attrition, observer effects
Observational studies
Difficulties because researchers must devise a way to systematically record amid chaos. Can use frequency method, duration method, intervals method (does behavior occur within pre-set time intervals)
